COMP61021 Modelling and Visualisation of High-Dimensional Data syllabus 2019-2020
A major component of machine learning and data mining is dealing with the high dimensional data that arises. Typical examples include pixels from an image (millions of dimensions), medical data bases (perhaps hundreds of dimensions, often with missing values), video clips and speech signals (time series data of very high dimensions), and gene expression data (expression values of many thousands of genes). Dealing with high dimensional data is a key challenge for modern computer science.
In this course we will consider how to develop appropriate algorithms for modelling and visualizing these high dimensional data sets and gain insights into these algorithms from theoretical and empirical perspectives. We will demonstrate how essential algorithms are derived in a step-by-step way as well as how important algorithms can be applied through the use of examples and real world problems. We will cover approaches from machine learning, statistics and neural computation in this advanced machine learning and data mining course unit.
This course unit aims to introduce students to state-of-the-art approaches to dealing with high dimensional data based on dimensionality reduction and provides experience of research such as literature review and appraising research papers in modelling and visualization of high dimensional data. In particular, transferable knowledge/skills, essential to original researches, are highlighted in this course unit.
- Mathematics Basics
- Principal component analysis (PCA)
- Linear discriminative analysis (LDA)
- Self-organising map (SOM)
- Multi-dimensional scaling (MDS)
- Isometric feature mapping (ISOMAP)
- Locally linear embedding (LLE)
three hours per week (5 weeks)
three hours per week (5 weeks)
In general, feedback is available for the assessed work.
For coursework, the feedback to individuals will be offered.
For exam, the general feedback to the whole class will be given in writing.
- Analytical skills
- Group/team working
- Oral communication
- Problem solving
- Written communication
On successful completion of this unit, a student will be able to:
- describe the curse of dimensionality and its implication in different learning paradigms including supervised and unsupervised learning
- understand the general motivation and main ideas behind dimension reduction techniques
- understand the advantages and the disadvantages of the learning algorithms studied in the course unit and decide which is appropriate for a particular application
- derive the principal component analysis (PCA) and the linear discriminative analysis (LDA) algorithms
- apply the learning algorithms studied in the course unit to simple data sets for dimension-reduction related applications
- implement PCA and Kohonen’s self-organised maps as well as apply them to real-world datasets for data modelling and visualisation
- evaluate the performance of the learning algorithms studied in the course unit and whether a learning algorithm is appropriate for a particular problem
- understand and appreciate main ideas underlying a state-of-the-art dimension reduction algorithm
|Introduction to machine learning||Alpaydin, Ethem, author.||9780262325745||The MIT Press||2014|
|Pattern recognition and machine learning||Bishop, Christopher M., author.||9780387310732||Springer||2006|
|Neural networks and learning machines||Haykin, Simon S., 1931-||9780131293762||Pearson Education||c2009.|
|Linear algebra and learning from data||Strang, Gilbert, author.||9780692196380||Wellesley-Cambridge Press||2019|
Course unit materials
Links to course unit teaching materials can be found on the School of Computer Science website for current students.